I chose this topic for my project because I wanted to use the same data set for RA and TS but, after much work, I am confused about the basic premise that I am testing. I originally proposed that more "current" stock prices (from Jan 1997 - Oct 2001) are predicted by the changing futures prices (independent variable), since the futures prices would show whether investors think the stock will go up or down. I began looking at this by using different lags to see if the stock price could be predicted based on, for example, a previous week's average futures contract. Now, after trying to do my analysis and rereading the introduction for this project, I'm thinking that I should be using the futures prices to predict only one stock price (December 2001). But then what would I be testing? It seems too simple that the futures contracts become a better gauge as time decreases.

I'm very confused but I don't want to change my topic since I've already put so much time into this one. Thank you in advance for any assistance you can give me.

[NEAS: You can use time series of the relation of forward prices to spot prices. For example, let F(t,k) be the forward price at time t of the stock price k months in the future, and S(t+k) is the stock price at time t+k. Fix k (say at three months), and F(t,k)/S(t+k) is a stationary time series. (You should test for stationarity using the methods in the course.) Financial economists argue about the mean of this time series and its ARIMA properties. You can evaluate the mean and fit an ARIMA process.

This is just an example. We do not grade the student project on its economic validity. We check if you properly use the methods learned in the time series course. For the time series above, Keynes discussed normal backwardation, and other economists have offered similar theories. Don’t worry about the theory. For the student project, get a stationary time series that is not a white noise process and fit an ARIMA model.]